MIT-licensed CLI + extension that gives DeepSeek Harness agents your real logged-in browser via a first-class dsh plugin injecting native browser_* tools and a live Web UI overlay.
DSH integration
Compatible
Author-claimed
Safety audit
Unaudited
Last verified
2026-08-21
License
MIT
01What can it help you accomplish?
Automate web tasks in your real, logged-in browser from DeepSeek Harness
Browser actions executed through native browser_* tools inside dsh, running in a dedicated Agent Window and reusing existing signed-in sessions without separate test accounts
DeepSeek Harness (dsh) users who need their agent to search, read and operate websites they are already signed into
Open, read and summarize web pages from a dsh prompt (AI web search workflows)
Page content fetched through your real browser and summarized in the agent conversation, e.g. `/browser-skill open example.com and summarize what is on the page.`
Anyone building AI web search or research workflows inside DeepSeek Harness who wants results from a real browser instead of a headless sandbox
Get past captcha, login and confirmation steps without breaking automation
Agent pauses at human-only steps, hands the browser over to you, and continues the task afterwards
Teams automating browser flows in dsh that include captcha, login or confirmation dialogs
02How to install into DeepSeek Harness
Prerequisites
- `bsk` CLI installed — macOS/Linux via the one-line install.sh script, Windows via install.ps1 in PowerShell (both install to `~/.local/bin`)
- BrowserSkill browser extension installed from the Chrome Web Store or Edge Add-ons (Chrome and Microsoft Edge supported)
Installation steps
- 01
Install the `bsk` CLI: `curl -fsSL https://raw.githubusercontent.com/Tencent/BrowserSkill/main/install.sh | sh` (macOS/Linux) or `irm https://raw.githubusercontent.com/Tencent/BrowserSkill/main/install.ps1 | iex` (Windows PowerShell)
$ curl -fsSL https://raw.githubusercontent.com/Tencent/BrowserSkill/main/install.sh | sh
- 02
Install the BrowserSkill extension from the Chrome Web Store or Edge Add-ons; on other Chromium-based browsers, install the Chrome Web Store build
- 03
Add the plugin to a dsh profile: `dsh plugin --profile web add @wxg-prc-cpg/browser-skill-dsh-plugin`
$ dsh plugin --profile web add @wxg-prc-cpg/browser-skill-dsh-plugin
- 04
Start that profile: `dsh --profile web` — the plugin carries its own copy of the skill, so `bsk install-skill` is not needed for dsh
$ dsh --profile web
Verify the integration
- Run `bsk --version` to verify the CLI binary is installed
03DSH integration and capability boundaries
First-class dsh plugin on npm (@wxg-prc-cpg/browser-skill-dsh-plugin) added to a dsh profile; injects native browser_* tools and a live Web UI overlay of each Agent Window
Native browser_* tools for dsh
a dsh profile with the plugin added via `dsh plugin`→injected browser_* tools (no shelling out to bsk) plus a live Web UI overlay of each Agent Window
The plugin invokes the local bsk CLI/daemon on your behalf to drive the browserReuse real login state
websites you are already signed into in your browser→the agent works with authenticated pages directly, without separate test accounts
If the agent needs a tab you already have open, it must borrow that tab explicitly and return it when the task is doneIsolated Agent Window execution
browser tasks issued by the agent→tasks run in a separate, visible Agent Window so you can keep using your own browser
Human-in-loop takeover
captcha, login, confirmation dialogs or other human-only steps→the agent asks you to take over and continues afterwards
04Who is it for? When not to use it?
Good for
- DeepSeek Harness (dsh) users who need their agent to search, read and operate websites they are already signed into
- Anyone building AI web search or research workflows inside DeepSeek Harness who wants results from a real browser instead of a headless sandbox
- Teams automating browser flows in dsh that include captcha, login or confirmation dialogs
Not for
- Only Chrome and Microsoft Edge are officially supported; other Chromium-based browsers are merely expected to work with unpacked extensions, and Firefox support is still planned.
- Even for DeepSeek Harness, the `bsk` CLI/daemon and the browser extension remain prerequisites — the dsh plugin alone is not sufficient to run.
05Compatibility, maintenance and safety notes
- Only Chrome and Microsoft Edge are officially supported; other Chromium-based browsers are merely expected to work with unpacked extensions, and Firefox support is still planned.
- Even for DeepSeek Harness, the `bsk` CLI/daemon and the browser extension remain prerequisites — the dsh plugin alone is not sufficient to run.
- The runtime covers macOS (Apple Silicon and Intel), Linux (x64 and ARM64) and Windows x64 only; other operating systems are not listed in the support matrix.
MIT · actively maintained (latest release dsh-plugin-v0.1.1, 2026-08-19)
06Frequently asked questions
How does BrowserSkill integrate with DeepSeek Harness — plugin or MCP?
It's a first-class dsh plugin on npm (@wxg-prc-cpg/browser-skill-dsh-plugin). Add it to a profile with `dsh plugin --profile web add @wxg-prc-cpg/browser-skill-dsh-plugin`, then start that profile with `dsh --profile web`. It injects native browser_* tools — no shelling out to bsk — plus a live Web UI overlay of each Agent Window.
What do I need installed before adding the dsh plugin?
The `bsk` CLI/daemon and the BrowserSkill browser extension are still prerequisites even for dsh. The plugin carries its own copy of the skill, so `bsk install-skill` is not needed — but the CLI and extension must be in place first.
Which browsers and operating systems are supported?
Chrome and Microsoft Edge are supported (other Chromium-based browsers are expected to work with unpacked extensions; Firefox is planned), on macOS (Apple Silicon and Intel), Linux (x64 and ARM64) and Windows x64.
Will the agent interfere with my normal browsing?
Browser tasks run in a separate, visible Agent Window so you can keep using your own browser. If the agent needs a tab you already have open, it must borrow that tab explicitly, return it when the task is done, and leave the rest of your browser alone.
What happens when a task hits captcha or login?
BrowserSkill has built-in human-in-loop: the agent asks you to take over the human-only step (captcha, login, confirmation dialog) and continues the task afterwards.
07Related DSH workflows
openviking
by volcengine
Self-evolving Context Database for AI Agents. Unify Agent Memory, Knowledge RAG and Skills.
colleague-skill
by titanwings
Turn memories and knowledge into warm, reusable agent skills — a skill generator for digital companions in the AI persona era.
archify
by tt-a1i
Agent skill for beautiful, verifiable architecture, workflow, sequence, data-flow, and lifecycle diagrams—self-contained HTML with motion and crisp export.
learn-harness-engineering
by walkinglabs
Harness engineering beginner tutorial, from 0 to 1
08Data and sources
BrowserSkill ships a first-class dsh plugin on npm as [`@wxg-prc-cpg/browser-skill-dsh-plugin`](https://www.npmjs.com/pa…
It injects native `browser_*` tools (no shelling out to `bsk`) and a live Web UI overlay of each Agent Window.
dsh plugin --profile web add @wxg-prc-cpg/browser-skill-dsh-plugin
This page is generated from the project’s public documentation, repository metadata and a structured parse of DSH Plugins; last verified on 2026-08-21. Found an error? Submit a correction.